Algorithm 10^310^4 10^5 10^6 10^7 10^8
Reinke0.04 1.2141.00- - -
- Reinke is the ``applyAt'' algorithm Claus Reinke posted here
while I'm not unhappy that this code is performing reasonably well, it was
mostly an attempt to get
Claus Reinke wrote:
folklore3: merging the sieves, instead of stacking them as
implicit thunks
Here's Claus's code:
primes3 = 2 : folklore3 [] [3,5..]
folklore3 pns xs = x : folklore3 (insert (x,x*x) pns') xs'
where (x,pns',xs') = sieve3 pns xs
sieve3 ((p,n):pns) (x:xs) | xn =
This isn't a variant of the folklore sieve, it's actually the real one!
:-)
well, one thing i was trying for (and what perhaps one might like to see in a
paper), is a way of relating the various code alternatives to each other, with
suitably similar intermediate stages. not quite as detailed
Claus, you're absolutely right that my paper could be improved in a
number of ways, and that there is actually a surprising amount of
meat on these bones for further analysis. We'll see what happens.
If it actually gets accepted for JFP, it'll undoubtedly finally
appear as a revised